Each image was transferred check details to a SIS AnalySIS FIVE database (Soft Imaging System GmbH, Münster, Germany). In addition, all data of the image analysis were entered into the program and stored in this database. The methods and results of the histopathological examination have been published elsewhere (Ernst et al., 2002, Ernst et al., 2005 and Kolling et al., 2008). The inflammation score, which was correlated with the genotoxicity markers, was based on a grading scheme comprising five degrees of alterations (0 = none; 1 = very slight; 2 = slight; 3 = moderate;

4 = severe; 5 = very severe). Grade 1 (very slight): <10% of lung tissue affected; For statistical analysis, the SAS software package (release 9.1 on Windows XP computer, SAS Institute, Cary, NC, USA) and Statistica (version 8.0, StatSoft Inc., Tulsa, OK, USA) were used. The image analysis data were imported into these software packages. P450 inhibitor Data were analyzed by using analysis of variance (ANOVA) as an overall test. If the group means differed

significantly by ANOVA, the treatment groups were compared with the control group using Dunnett’s test. The Tukey HSD test was used as another post hoc test for comparison among the different treatment groups, as this test is not based on comparison between treatment and control. Statistical significance was reached if p ≤ 0.05. Data were considered highly significant if p ≤ 0.01 or p ≤ 0.001. The data for evaluation of possible correlations between genotoxicity marker expression and, for example, histopathology or BAL data were extracted from the interim and final reports of the research projects of the German Federal Environment Agency (Ernst et al., 2002, Ernst et al., 2005, Kolling et al., 2008 and Kolling et al., 2011) and from the raw data of individual animals of the research projects. Correlations between the genotoxicity markers and other study parameters such as inflammation score and enzymatic activities or cell Amylase counts in BAL fluid were calculated

using the respective group mean values. In case of the inflammation score, the individual animal data were additionally used for determination of correlations, because the same animals were investigated for both genotoxicity marker expression and histopathologic evaluation of lung inflammation. The method of linear regression/Pearson product-moment correlation (SAS [Cary, NC, USA] software package Statistica or SigmaStat 3.1) was used to calculate the correlation coefficient (r) and the significance of the correlation (p-value). Correlation coefficients lacking statistical significance were rated as “correlation without significance” if r > 0.5. PAR synthesis was determined in particle-exposed lung tissue as a general marker for genotoxic stress. Three months after the first and one month after the last exposure, alveolar lining cells of quartz DQ12-treated rats showed a statistically significant, about 1.